Jump to content
Sign in to follow this  
Guest vlada stoje

1s global model from ASTER is FREE!

Recommended Posts

Guest cbuchner1
1s global ASTER dem is available for free
Excellent. If one could come up with an automated downloader and conversion tool for FS9 and FSX this would be awesome. A select-by-continent or select-by-region feature would be nice.Christian

Share this post


Link to post
Share on other sites
Guest vlada stoje

Hi Christian,still no success with the download:-) The system is limited at 100 1deg tiles for now.CheersVlada

Share this post


Link to post
Share on other sites
Guest cbuchner1
Hi Christian,still no success with the download:-) The system is limited at 100 1deg tiles for now.CheersVlada
Once you succeed, please let me know if the DEMs are in a suitable format for the FSX SDK's resample tool.Christian

Share this post


Link to post
Share on other sites
Guest vlada stoje

Hi Christian,the japan GDEM site at http://www.gdem.aster.ersdac.or.jphas quite simple system, but I was unsuccessful there. The NASA sitehttps://wist.echo.nasa.gov/~wist/api/imswelcomein Sioux Falls has their standard complicated system with many stepsto push the order to shopping cart etc. I tried just small area N51E014 toN49E014 and got the ftp pull address with the finished orderftp://e4ftl01u.ecs.nasa.gov/PullDir/0301813176ZCPiU.zipthe archive with 3 tiles in geotiff format is about 25 MB.Comparison with SRTM is in this animated gifhttp://www.volny.cz/stoje/vera/srt_gdm1.gifsrt_gdm1.gifit should change from SRTM to ASTER GDEM.I like the data, and sure the new areas in latitudes 60

Share this post


Link to post
Share on other sites

Hi Vlada.The data I've looked at still has some clouds not removed, and the sea level determination is not very good.But, I think we could come up with a way to "correct" the data.If we were to make a 3 arc second, one degree, SRTM3 tile into a 1 arc second tile, we'd have a comparison to an acceptable set of elevations. Then a program could compare the two, using ASTER elevations for the result, unless those elevations exceed a +- 100 meter difference. Those areas could be voided, with SRTM3 values substituted. That would get rid of the remaining clouds, holes and spikes. I think you already experimented with creating the 1 arc second tiles from 3 arc second data... this would work here.Also, We could change the sea level to 1 meter, and then overlay the SWBD waterbody flattens over the DEM to get a better sea level determination back to zero elevation. The SWBD does have a few mistakes, but those could be corrected at that time as well.The end result would be a cleaned version of the ASTER data, that should be better than the SRTM, with the SWBD used for sea level.Dick

Share this post


Link to post
Share on other sites
Guest cbuchner1
Hi Vlada.But, I think we could come up with a way to "correct" the data.
I recommend to use a GPU for so much raw data processing, so it does not take another year ;)It will also need several terabytes of storage to do this thing globally on raw data. Ouch.So how will one person be able to download everything? Should we mount a collaborative downloadeffort?

Share this post


Link to post
Share on other sites
So how will one person be able to download everything? Should we mount a collaborative downloadeffort?
I'm getting the majority of Alaska downloaded today. Will have to start looking at files to see what quality is involved. Dick's posting I understand in concept, but not the implementation.

Share this post


Link to post
Share on other sites
I'm getting the majority of Alaska downloaded today. Will have to start looking at files to see what quality is involved. Dick's posting I understand in concept, but not the implementation.
What I proposed to Vlada is using ASTER and SRTM to produce an improved dataset. I think the tiles are 3601x3601 points for ASTER per degree squared tiles. The SRTM 90 meter data would have to be expanded to 3601 from 1201 points to match the same data points. But this requires a cleaned and acceptable SRTM set ( maybe CGIAR data? ) The US has NED 30 meter data that wouldn't need to be replaced, and Canada may also have a better data set than we could derive. But this should work for the rest of the world.Using a comparison of the 16-bit signed data, if the ASTER point is within 100 meters... or perhaps a percentage amount ( say 10% ) then the ASTER point is binary written to a new datafile; else the SRTM would be written to the new datafile. I would also fill any remaining SRTM voids with the ASTER elevation ( but with CGIAR there shouldn't be any ). This should get rid of ASTER cloud errors, and correct ASTER spikes, holes and voids with finished SRTM data on a point by point basis for each tile.So we'd use set A, compared to set B to derive set C. As long as A is OK compared to B, we use A to make C... in bad areas of A, we'd fill with B data.And I would write all 0 meter data as 1 meter in the new datafile... so the SWBD shapefiles could be applied to the new tile, giving us a better zero meter sea level.At that point we'd have a corrected ASTER 30 meter HGT file ( 16-bit signed binary ) ready to go for FS.I understand there may be a 1/2 pixel shift in the CGIAR data... but that may actually help Vlada, as he already has an algorithm to create 30 meter data from 90 meter data using just this pixel shift. Alternately, exporting the CGIAR data as 30 meter spacing from Global Mapper as 16-bit binary, as well as exporting the ASTER as 16-bit binary may be good enough to produce the two input data sources.The final tile may still have odd anomalies such as "mole runs"... odd lines through the data that would not be caught by the SRTM point comparison. A version 2 of this new tileset, would have to have manually voided areas, backfilled by SRTM, to eliminate further errors, such as the mole runs.If we don't do something like this, it will be 4-5 years before the ASTER is "finished", and it may also get dumbed-down to 90 meter data, just as the SRTM world set was.Dick

Share this post


Link to post
Share on other sites
Guest vlada stoje

Hi Dick,yes combining of sources is good way, for example Jonathan Ferranti used for the High Tatras very similar method as you described

Share this post


Link to post
Share on other sites
Guest cbuchner1
The final tile may still have odd anomalies such as "mole runs"... odd lines through the data that would not be caught by the SRTM point comparison. A version 2 of this new tileset, would have to have manually voided areas, backfilled by SRTM, to eliminate further errors, such as the mole runs.Dick
The mole runs could get caught by running some analysis on a difference image between SRTM and ASTER. If the mole runs are mostly straight lines, a line detection algorithm such as a Hough Transform might just work fine. This could automate the detections of such artifacts.

Share this post


Link to post
Share on other sites

Interesting datasets, not so good for Europe or the USA/Canada but for some locations in Australia, SE Asia, Africa and South America it looks very good for some local area features that SRTM doesn't tend to cover...I like the sound of Dick's idea for doing analysis and comparisons, but wouldn't there tend to be a 'stair step' effect in between the two datasets, i.e. say one dataset is RMSE +20 while the other is RMSE -20, that would end up with a 40m 'step' between the datasets... Unless the software can blend the datasets together to remove any such seams...


Dean Mountford
Ultimate VFR

Share this post


Link to post
Share on other sites
Interesting datasets, not so good for Europe or the USA/Canada but for some locations in Australia, SE Asia, Africa and South America it looks very good for some local area features that SRTM doesn't tend to cover...I like the sound of Dick's idea for doing analysis and comparisons, but wouldn't there tend to be a 'stair step' effect in between the two datasets, i.e. say one dataset is RMSE +20 while the other is RMSE -20, that would end up with a 40m 'step' between the datasets... Unless the software can blend the datasets together to remove any such seams...
Hi Dean.I think the CGIAR srtm-based data was reworked by converting the mesh to contours, then adjusting those contours before converting them back to a dem. Maybe that would help.In order to prevent the stair-step, you'd have to find the anomaly, then find a nearly matching contour in the patching dataset, then create a void covering the anomaly that extends to the nearly matching contour by deleteing those contours. Then add in the new contours from the patching set.Alternately, just void the anomaly, add the new contours, and adjust one or two contour lines to blend the transition. I don't think anything automatic will work here, and the world is a large place.Dick

Share this post


Link to post
Share on other sites
Guest cbuchner1
I think the CGIAR srtm-based data was reworked by converting the mesh to contours, then adjusting those contours before converting them back to a dem. Maybe that would help.
I think this will require some complex mapping tools. Generating contours from DEMs and vice versa is no simple maths. And doing this on entire continents would probably take weeks.

Share this post


Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now
Sign in to follow this  

  • Tom Allensworth,
    Founder of AVSIM Online


  • Flight Simulation's Premier Resource!

    AVSIM is a free service to the flight simulation community. AVSIM is staffed completely by volunteers and all funds donated to AVSIM go directly back to supporting the community. Your donation here helps to pay our bandwidth costs, emergency funding, and other general costs that crop up from time to time. Thank you for your support!

    Click here for more information and to see all donations year to date.
×
×
  • Create New...